Televised Presidential Debates and Its Influence on Voters
In debates, candidates must face unanticipated questions forcing spontaneous responses. Voters obtain a less artificial view of the candidates in debates than from speeches or television spots. Speeches often go through many drafts while television spots have multiple takes (Benoit, et. al., 2002).
Issue positions and candidate images are two important predictors of how people will vote. The authors asked open-ended questions based on National Election Studies data about presidential candidates and concluded that three groups of comments emerged: comments about a political party, comments about a candidates’ issue position, and comments about a candidate’s personality traits. The author explains that this pattern has continued for over twenty years (Mille, et. al., 1984).
Some scholars say that issue learning and image formation comparisons are like comparing apples and oranges – two different concepts that are incapable of any valid comparison but an indirect comparison can be made. Two research questions can be posed: whether the debate causes change in images, and whether the debate causes an improvement in issue knowledge (Zhu, et. al.1994). A relational comparability approach is used for two concepts that cannot be compared directly. These researchers recommend comparing exposure to mass media and its effects on interpersonal communication by exploring the ability to predict knowledge, attitudes, or behavior based on these information channels (Chafee and Mutz, 1988). Zhu, et. al., explain that televised debates are meant to contribute positively to the democratic process. Debates make negative contributions to this process when they result in audiences learning more about the candidate’s personalities than about their issue positions (Zhu, et. al., 1994).
The issue-only school of thought predicts that televised debates will enhance issue learning but not image formation. Audiences evaluate candidates based on their partisan beliefs resulting in the inability of the debate to affect images (Bothwell and Bringham, 1983; Hagner and Rieselbach, 1980; Kraus and Smith, 1962; Payne, et. al., 1989; Tannenbaum, et. al., 1962). Debates can only reinforce the preexisting images of candidates. These preexisting images are created through stereotyping theory in which audiences use categorization as a cognitive shortcut to evaluate candidates. These cognitive categories are resistant to change. (Bowes & Strentz, 1979; Carter, 1962).
There are mechanisms that audiences use to gain information about candidate issue positions. Debates are major media events that bring campaign issues to the forefront and define what positions each candidate has (Drew & Weaver, 1991). Uses and gratification theory explains the knowledge gap that exists between viewers and nonviewers of debates (Bishop, et. al., 1980).
The image-only school predicts that audiences will receive little gain in issue knowledge after a debate but they will be impacted greatly by the candidates’ images. Even though television conveys both verbal and visual messages, it essentially is a visual message medium requiring less involvement by audiences to process the message than that of processing messages in print media (Graber, 1987). Candidates’ visual images contribute to their personality images while verbal messages contribute to substantive images. Television excels in describing policy problems but fails to explain possible solutions (Hellweg et. al., 1992).
Zhu, et. al., examined 36 studies of presidential debates that used debate exposure as an independent variable and categorized issue learning and image formation as dependent variables. Most of these studies were done either on a group of local adults or college students. Only four of the studies used national general public samples. Most of the studies looked at debates during the general election campaign and examined the first debate. The first debate has the greatest impact than that of later debates (Zhu, et. al., 1994). Candidates who are successful in first debates usually, “direct their remarks at highly targeted audiences, develop an overall theme throughout the debate, debate not to lose by avoiding specifics and making use of proven safe responses, present themselves as vigorous active leaders, foster identification of themselves with national aspirations, foster identification of themselves with the dominant political party/philosophy, and personify themselves as exemplifying a desirable characteristic” (Friedenberg, 1997).
Pfahu argues that certain dimensions enable a candidate to be successful in a debate. These dimensions include categories that he defined as: “immediacy/affection, which consists of involvement, warmth, enthusiasm, and an interested in the receiver; receptivity/trust, which comprises sincerity, honest, interest in communicating, and willingness to listen; similarity/depth, which includes friendliness and caring; and lesser dimensions (Burgoon and Hale, 1984 and 1987). Clinton relationally connected with viewers in the 1992 presidential debates especially during the Town Hall debate format. His friendliness, sincerity, interest in communicating, and willingness to listen enabled him to win in the debates especially since his opponent; George Bush made a nonverbal mistake in the debate that was looking down at his watch (Pfau, 2002).
Twenty-seven of the studies looked at by Zuh used surveys to collect data; six had experimental designs with a controlled or manipulated variable, while only three used a random procedure to assign subjects to groups (Hawkins, et. al., 1979, Lemert, et. al., 1983, Pfau, 1987). Two studies used features of both survey and experimental design. One third of the studies did not measure whether the audience watched the debate. Some studies assumed that the debate was the only source of the change or lack of it by comparing issue/image scores that were measured before and after a debate. The two points of measurement were one week or more apart (L. Becker et al., 1979; Hagner & Rieselbach, 1980). Most of the studies had a pretest a few days before the debate to offset any sensitization effect and a posttest immediately after the debate to offset postdebate political learning. However, the participants could have been influenced by other sources of campaign information between the two tests. This is what Zhu calls a “contaminated learning” effect. Only seven studies controlled for this effect by comparing a control group of non-viewers with an experimental group of viewers (Zhu, et. al., 1994).
Issue learning is divided into seven categories: issue knowledge, issue opinion, issue congruency, issue discrimination, and issue salience. Issue knowledge is how the audience judges a candidate’s issue position. Issue opinion are the audience’s subjective views on issues. The similarity of audience’s position on issue and a candidate’s position is shown in issue congruency. Issue discrimination is the differences between perceived positions taken by multiple candidate. The audience’s issue salience is how important the audience sees an issue. The candidate’s issue salience is the audience’s opinion on how important a candidate perceives an issue (Zhu, et. al., 1994).
Issue perception is divided into four categories: image congruency, issue discrimination, and image salience. Image congruency is the differences between the perceptions of a candidate and that of an ideal president. Image discrimination is the differences between traits of the candidates. Image salience is how traits of a candidate influence one’s vote choice (Zhu, et. al., 1994).
The four measures that stood out as the most frequently used were image perception in twenty-four studies, issue discrimination in nine studies, issue knowledge in eight studies, and image discrimination in six studies. Studies of debate effects in the future should be a basic design contrasting viewers and nonviewers with tests both predebate and postdebate. There also should be two groups – a forced viewing group and a natural viewing group. The natural viewing group would help offset laboratory affects (Zhu, et. al., 1994).
Zhu, et. al., performed a study on the first 1992 presidential debate between George Bush, Bill Clinton, and Ross Perot on college students at the University of Connecticut using their recommendations and knowledge from past studies. It was not a random subject study because all of the students were enrolled in introductory classes in communication and public speaking. Studies done on college students are not accurate of the population as a whole. For example, 32% were “very interested” in national politics as compared to 49% in the electorate national sample. The questionnaire dealt with questions about the perception of candidates’ characteristics, candidates issue positions, salience rating of election issues, the performance of the candidates on these issues, and the personal background of the respondent The last two affects could not be measured accurately in this study (Zhu, et. al., 1994).
The results concluded that those who watched the debate knew 34% more about Bush’s issue position, 24% more about Clinton’s, and 39% more about Perot’s. It is interesting to note that knowledge of debated issues has a 0.109 correlation coefficient to employment. In other words, students who have a full or part time job had more knowledge about the issues. A 0.189 correlation coefficient was found in students who had high interest in politics. This means that students highly interested in politics also knew more about the issues, which makes sense. The study answers its initial question by stating that the first presidential debate in the 1992 presidential election had a large effect on audiences issue knowledge. However, it had no impact on the perception of candidates personalities. The least known candidate, Perot, was the only candidate that researchers found audiences learned most about during the debate while they learned the least about the two candidates, Clinton and Bush, who were the most known. A debate does not have the potential to change an audience’s focus from issues to personality because the public sees a candidate’s ability to debate and their ability to be a president as two distinct entities (Zhu, et. al., 1994).
Audiences may know more about the issues after debates, but sometimes debates do not drastically change which candidate the public favors. In the Gallup poll in the week before the first debate, 52.42% of voters favored Clinton and 36.85% favored Dole. In the week after the first debate, 52.71% of voters favored Clinton, a 0.29% increase, and 34.14% favored Dole, a 2.71% decrease. Although there are differences in the percentages, they are two small to attribute to the debates. Other polls released at the same time were consistent with these Gallup poll findings (USA Today, 1996). Friedenberg concludes that Clinton debated effectively and had a large lead ten days before the election while Dole debated less effectively but he managed gained ground after the debates, but not enough to close the gap between him and Clinton (Friedenberg, 1998).
Benoit, et. al., argue that in some ways primary debates are more important than general debates. They allows partisan voters to decide who will represent their party in the upcoming election, it narrows the field of potential candidates, and it determines the outcome in the general election when a candidate competes with a vulnerable incumbent. Incumbents have an advantage in most elections but several have still lost the election such as Gerald Ford in 1976, Jimmy Carter in 1980, and George Bush in 1992. Primary debates are watched by fewer voters but they are generally targeted to smaller audiences and allow voters to see multiple candidates at once. In 1992, there were 13 primary debates and 3 general debates so primary debates allows more opportunities to influence voters (Benoit, et. al., 2002).
Benoit conducted a study on issue knowledge, candidate evaluation, and vote preference in the 2000 presidential primary debates. In the Republican Primary debate study, thirty-four participants had not watched a Republican Primary Debate before while in the Democratic Primary debate study, seventy-eight participants had not watched a Democratic Primary Debate before (Benoit, et. al., 2002).
Study one done on the Republican Primary Debate concluded that the debate did create impressions of candidates for voters who had not yet formed an impression of a candidate thus confirming hypothesis 1. On the pretest, approximately thirty-seven participants indicated on their form that they “didn’t know” which candidate they favored while on the posttest, approximately three participants indicated the same. Approximately thirty-four participants made a voter preference after watching the debate. In a similar study done on the Democratic Primary Debate, about ten participants marked “don’t know” on the pretest while only two participants marked “don’t know” in the posttest. Eight participants knew their voter preference after watching the debate (Benoit, et. al., 2002).
Hypothesis 2 stated that watching a primary debate changes evaluation of candidates’ character. This hypothesis was concerning voters who already formed an impression of a candidate’s character. The hypothesis was unable to be tested in the Republican Primary Debate because too many respondents reported saying they “didn’t know” whether the debate changed their impression about the candidate. The sample size of thirty-four proved to be inadequate in proving or disproving this hypothesis. In the Democratic Primary Debate, the hypothesis also was unable to be tested but after the debate respondents viewed Gore as more honest, experienced, fair, competent, decisive, attractive, trustworthy, energetic, and understanding than before (Benoit, et. al., 2002).
Hypothesis 3 stated that watching a primary debate would enhance viewers’ knowledge of the policy positions of the candidates. This hypothesis was supported by the fact that participants in the Republican Primary Debate study had more correct answers to policy position questions in the posttest with 7.9 questions answered correctly than the pretest with 1.1 questions answered correctly. This means that on average participants answered 6.8 more questions correctly because of the debate. Results from the Democratic Primary Debate study concluded that only 1.1 respondents answered policy questions correctly in the pretest while in the posttest 5.6 respondents answered questions correctly. This means that on average 4.5 respondents gained policy position knowledge because of the debate (Benoit, et. al., 2002).
Hypothesis 4 predicts that watching a primary debate can change a viewer’s voting preference. This hypothesis was correct in the Republican Primary Debate study. Significant changes were found in vote choice after viewers watched the debate. However, in the Democratic Primary Debate study, there was no net shift favoring either Bradley or Gore with fifteen participants changing their preference from Gore to Bradley and fourteen participants changing their preference from Bradley to Gore. This means twenty-nine participants changed their voting preference after watching the debate (Benoit, et. al., 2002).
Hypothesis 5 states that watching a primary debate changes the confidence level in vote preference. This hypothesis was true in the Republican Primary Debate study. Debate viewers were asked to rate confidence in their voting preference from 0-100, with 1 being very uncertain and 100 extremely sure. The average pretest score was 38.9 but in the average posttest score was 52.9, a change of 14 points. Likewise, in the Democratic Primary Debate study participants’ average pretest score was 38.2 and their average posttest score was 69.6. The difference between the pretest score and the posttest score is 31.4 points (Benoit, et. al., 2002).
Hypothesis 6 predicts that watching a primary debate changes the level of confidence for viewers who did not change their vote choice. This hypothesis was true in the Republican Primary Debate study. Ten participants out of thirty-four did not change their voting preference after the debate but all ten changed their level of confidence with seven having an increase in confidence and three having a decrease in confidence. In the Democratic Primary Debate study, forty-eight out of seventy-eight participants did not change their vote choice after the debate but forty out of forty-eight changed their level of confidence. Thirty-five increased confidence in their candidate, six decreased confidence, and eight reported no change in confidence (Benoit, et. al., 2002).
Voters cannot use political party as a basis for their vote choice in primary campaigns since several candidates from the same party run against each other (Benoit, et. al., 2002). Most voters change their partisan choice each election based on their positions on the issues and their feelings on the abilities of the candidates (Pomper, 1975). These hypothesizes show that viewers form impressions of the candidate’s character, more significant changes were found in the evaluations of the Democratic Party Debate candidates, Bradley and Gore, than in evaluations of the Republican Party Debate candidates, Bauer, Bush, Forbes, Hatch, Keyes, and McCain. They also show that viewers of both Republican and Democratic debates gained knowledge about policy issues from watching the debates and that there was a large change in vote choice after watching the Republican Party Debate. Over one third of those watching the Democratic Party Debate changed their voting preference. Finally, the conclusions show that viewers of both the Republican Party Debate and the Democratic Party Debate had greater confidence in their vote choice (Benoit, et. al., 2002).
Some researchers have claimed that presidential debates only inform those who are already informed and further widen the knowledge gap. They explain that if there were no debates, these viewers would have learned the same information from other sources (Bishop, et. al., 1980, McLeod, et. al., 1979). Previous studies have lumped debates together and did not distinguish between debate-only viewers and those who watched both a debate and the news special following it. However, in the study done by Lemert, debate exposure resulted in knowledge for both of the post debate samples in his experiment. Two of the three 1988 presidential debates lead to higher knowledge scores for people watching the debate compared to those who did not watch the debate. The results also shows that people who watched both the debate and the following news show always had higher knowledge scores than those who just watched the debate. The only exception was the Quayle-Bentsen debate. Quayle’s competence was questioned in this debate and there was little policy discussion (Lemert, 1993). In this debate, the words “qualifications” or “experience” were used twenty-seven times. Hershey’s study concluded that more policy content was found in the other two 1988 presidential debates (Hershey, 1989).
A study was conducted by Tsfati looking at the effect the 1996 presidential debate had on audiences, he concluded that out of those who watched the debate, 70% talked about it with others. Only 40% of those who watched only the news coverage of the debate talked about the debate with others (Tsfati, 2003). He further explains “exposure to the debate and its news coverage are not associated with predate partisanship and candidate preferences when controlling for demographic and political involvement variables” (Tsfati, 76, 2003). Political involvement is the main factor determining whether a person chooses to be exposed to the debate and its news coverage. This makes practical sense because audiences do not know ahead of time which candidate will do better or how commentators will react to the debate.
Hypothesis 1 in Tsfati’s study stated that pre-debate attitudes toward the candidates would influence an audience’s debate evaluations. This hypothesis proved to be correct even though participants were asked to evaluate each candidate regardless of which candidate they support. This relates to the stereotyping theory discussed earlier.
Hypothesis 2 stated that audience evaluations regarding the debate would be affected by general campaign involvement but this hypothesis could not be supported.
Hypothesis 3 said that respondents would be influenced by their exposure to news coverage of the debate and that would affect who they would claim to be the winner of the debate. The hypothesis was correct and the study shows that the more debate news coverage the participant received, the more likely he or she would say that Clinton won the debate. This is because many commentators stated that since Clinton controlled the agenda and remained calm he did a better job than Dole who got angry and attacked his opponent.
Hypothesis 4 predicted that debate conversations would affect debate evaluations. The hypothesis was proven to be true. “Each one-unit increase in the debate-discussion variable was associated with a 174 percent increase [100(e1.01-1)] in the odds saying that Clinton had done better” (Tsfati, 78, 2003). This means that one’s interpersonal environment influences one’s perceptions of who had done a better job in the debate.
Hypothesis 5 predicted that a relationship between the amount of debate exposure and the extent of exposure to debate news coverage. This hypothesis was correct which means that as respondents watch more of the debate, the effect of news coverage becomes less. News coverage has no affect on individuals who watched the entire debate (Tsafti, 2003).
The 2000 presidential debates had a significant impact on the subsequent election, possibly to the degree the 1960 debates between Richard Nixon and John F. Kennedy had on that election. The Commission on Presidential Debates (CPD) determines rules of general debates and who can participate. In the 1992 presidential election, third party candidate, Ross Perot, participated. However, the CPD later determined that general debates should only include those candidates who have at least 15% support from the national electorate using an average of five polls including the ABC News/Washington Post Poll, CBS News/New York Times Poll, NBC News/Wall Street Journal Poll, CNN/USA Today/Gallup poll, and the Fox News/Opinion Dynamics Poll. This new requirement resulted in angering third-party candidates who would not be able to generate that much support but wanted the opportunity to speak on television through this format (Friedenberg, 2002).
The two major candidates in the 2000 election, Al Gore and George W. Bush, agreed to the format of three presidential debates, each to be 90 minutes long. The candidates would answer moderator questions while standing behind a podium for the first debate. In the second debate, candidates would be seated at a table with the moderator who would pose questions to the candidates. A town meeting format, a format more informal than the first two, would be used for the third debate in which candidates would sit on stools and answer moderator questions but would be free to move around. The questions posed by the moderator in this debate would be questions submitted by the audience members. Audience members would consist of undecided voters and would be chosen by the Gallup polling firm (Friedenberg, 2002). The three debates offered three different formats and a vice presidential debate occurred between the first and second presidential debates. Jim Lehrer, host of the Public Broadcast System’s nightly news show, would serve as moderator for all three of the debates (Balz, 2000).
The Gore campaign set high expectations for the debate because Gore was known as a debate guru and it helped lower expectations for Bush. Two of the three debate formats were formats that favored Gore’s formal approach and only one of the formats favored Bush’s informal approach. Gore’s exaggerations helped gain support for Bush because people began to question the honest and credibility of Gore. Bush’s lack of debate and verbal skills ultimately lowered expectations for his performance in the debate. In the debates, Bush was more direct and specific answering the moderator’s questions while Gore failed to completely some questions he was asked, repeatedly disregarded time limits in the debate, and did not have the ability to answer a simple yes/no question without going into a short explanation (Friedenberg, 2002). According to the CNN-Time poll taken in 2000, “only 54 percent of the public felt Gore was trustworthy enough to be president, and 60 percent felt that he would say anything to get elected president” (Lester, 2000). 66% of voters polled said Bush was trustworthy enough to be president while only 40% felt Bush would say anything to win the election. Concerns about Gore’s exaggerations and credibility were not readily apparent after the Democratic National Convention but emerged after the first debate (Lester, 2000).
Forty-five undergraduate students watched the third presidential debate of election 2000 and completed a Policy Views Questionnaire before and after the debate. A repeated-measures t test showed a significant difference between pretest and posttest means of t43=2.76, p<0.08. The pretest mean was 10.0 with a standard deviation of 3.3 while the posttest mean was 10.9 with a standard deviation of 3.1. At pretest, participants were more likely to vote for Gore but watching the debate raised the support. This study has no control group, small sample size, and it was not a randomized sample demographically. It does help show that debates do change a viewer's attitude (McIntosh, 2002).
Both general and primary presidential debates have an impact on voters. Political debates force candidates and voters to address important issues and engage in real political dialogue. If there were no face-to-face debates, candidates would repeat themselves, not be forced to defend their positions, and not be forced to refine their political thinking. He encourages the government to provide public resources to all candidates so that it can truly be used as a designated public forum. Unless the government shows a compelling case against it, third-party candidates should be included in all presidential debates. Political debate and discourse is severely limited when rules state that certain candidates cannot participate in what should be a public forum. He goes on to say that if a debate format cannot reasonably accommodate additional candidates, than additional debate should be held to accommodate all candidates equally (Raskin, 1999).
These debates help voters see the policy positions the candidates are taking and enable them to vote more intelligently. Media coverage of debates enables important information in the debate to be presented to those who did not watch the debate. Although those who watch news coverage of a debate, but not the debate itself, are more likely to agree with how the media perceives the election, the debate and the coverage about it encourage Americans to vote. Without the attention from the media about the debates, the debates probably would have little influence on voters because voters usually get their cues on what is important by what is featured on the news. The core of democracy is based on the premise that ideas are presented in an open forum and that premise is one of the purposes of presidential debates.
